Global trend assessment of land use and land cover changes: A systematic approach to future research development and planning

2.1 Bibliometric data processing and PRISMA approach

Scopus, ScienceDirect, and WOS databases were utilised to generate bibliometric data from 1993 to 2022. Similar studies have widely used multiple databases for systematic review and bibliometric analysis (Afuye et al., 2021b; Macarringue et al., 2022). This study adopted the PRISMA technique to comprehensively review LULC change studies (Velastegui-Montoya et al., 2022). The search key terms “Land use and Land cover change” OR “Land-use and Land cover change” OR “Land use/Land cover change” OR “Land use and Land cover changes” were used to retrieve all open access and English-written published documents within a niche area. Data screening, cleaning, and processing were applied to remove duplicates and irrelevant documents, reducing the vocabulary dimension (Nduku et al., 2021). The data screening involves skimming and scanning documents using search key terms, including article titles, abstracts, and keywords. It removes irrelevant documents and ensures that the final count of published papers in an Excel spreadsheet matches the search terms. This process helps clean and process irrelevant documents and ensure the final count’s accuracy. The PRISMA tool, Zotero (v6.0.7), RStudio (v4.2.3) and Biblioshiny software were used to filter non-informative words and collate all retrieved published documents (Aria et al., 2022). Open-source environment tools like PRISMA, Zotero, and RStudio provide an easy-to-use comprehensive platform for bibliometric data (Aria et al., 2022). PRISMA offers an overview guide to selecting documents from databases to the point of compilation for analysis. R Studio open-source packages, biblioshiny, and RStudio software facilitate script creation, merging multiple datasets, and removing duplicates. Zotero allows for validation before analysis and visualisation of the entire document collection of bibliometric data.

The study analysed 2710 published articles using bibliometric tools and a web interface to evaluate annual scientific production growth trends, global scientific contributions per article, and citation density. The PRISMA stages were chronological, starting with finding published articles using search keywords. Data screening was performed to identify eligible articles, followed by data processing to merge retrieved articles from Scopus, ScienceDirect, and WOS databases. Data cleansing was performed to remove duplicates and prepare documents for analysis. The density visualisation map of the top author's keyword terms was created using VOSviewer (v1.6.20) and the Research Information Systems (RIS) file format for bibliometric data. The minimum number of keyword occurrences was randomly selected at 50 to achieve high-frequency keywords in LULC change studies. The study measured the average growth rate (AGR), the actual difference between previous and current years, and the number of publications in annual scientific production growth (Ruiz-Rosero et al., 2019) using equations (1) and (2). The citation analysis was retrieved through a built-in formula on the R package repository with the following equation (3). A density visualisation map for top authors' keywords was generated on VOSviewer software.

Formula description: $A$ is a document $\times$ cited reference matrix ${(A}^{\prime })$ and $B_{\mathit{coup}}$ is a non-negative and symmetrical matrix. The PRISMA approach below was used for document selection criteria as shown in Fig. 1.

Close

Fig. 1

The PRISMA flowchart for document selection criteria.

Export to PPT

3.1 Scopus, ScienceDirect and WOS extracted bibliometric information

The study analysed 2710 LULC studies from Scopus, ScienceDirect, and WOS databases during the survey period, as shown in Table 1. The bibliometric analysis revealed 1736 articles, 148 proceeding papers, 70 reviews, 16 book chapters, 740 other document types, and 834 sources of journals and books. During the study period, 9999 authors and 4.9 co-authors contributed to strengthening scientific research publication developments on LULC changes. International co-authorships and single-authored documents contributed 1.73 % and 101 authors to published documents in LULC change studies. Between 1993 and 2022, global research on LULC experienced an average annual growth rate of 43.51 %, indicating increased attention to monitoring change detection in recent decades.

3.2 Annual scientific trend dynamics on land use and land cover change studies

Fig. 2 shows an annual scientific trend on LULC change studies, with an annual scientific production growth rate of 24.37 %. This indicates consistent global advancement in research outputs, as recorded in published documents. Between 1993 and 2009, there was a low progression in published research output. However, an exponential trend began in 2010, peaking in 2022, indicating more detailed advances in LULC change research. The study reveals significant exponential growth in LULC change studies from 2019 to 2022, aligning with Price law in bibliometric analysis (Price, 1976). This growth is attributed to the high demand for agricultural and habitable land use between 2021 and 2022, with 2022 recording the highest number of publications in this field. The United Nations Environment Programme (UNEP) is working towards sustainable management and restoration of Sustainable Development Goals (SDG15): Life on Land, aiming to ensure the benefits of land-based ecosystems, including sustainable livelihoods for generations (UNEP, 2020). SDG 15 aims to prevent desertification, enhance sustainable practices, reverse biodiversity loss, and manage forest resources effectively to prevent land degradation and biodiversity loss (UNEP, 2020). The research on LULC change studies has gained global attention, contributing to a temporal trend in scholarly output since 2000 (y = 13.078e^26.164x), as shown in Fig. 2.

Close

Fig. 2

Annual scientific production growth trend from 1993 to 2022.

Export to PPT

3.3 Global scientific contributions per article and citation

The research on LULC changes was strengthened by article publications and citations from single and multiple countries as shown in Table 2. China was the most influential country (N = 433), followed by the USA (N = 273) and India (N = 120). The total number of publications, total citations, average article citations, single-country publications (SCP), and multiple-country publications (MCP) varied across countries. The USA (TC = 12153 AAC = 44.50) accounted for the most total citations during the survey period, followed by China (TC = 8628; AAC = 19.90) and Germany (TC = 8628; AAC = 19.90). These countries significantly contributed to SCP and MCP articles in single and joint-authored articles. The results indicate that developed countries, such as Brazil, Indonesia, and Ethiopia, made significant contributions to publications and citations compared to developing nations. The study revealed significant disparities in multiple-country publications in developing nations, primarily due to language barriers, research grants, and autonomous research (Huang et al., 2016). The research field of LULC change is driven by factors such as intra-country contribution and low inter-country international co-authorships, which have a significant effect. The significant effect of the intra-country contribution and low inter-country international co-authorships of 1.73 % were driving factors of LULC change research field. During the survey period, the study revealed that the lack of research grants significantly impacted the collaboration and productivity of researchers across different countries.

3.4 Institutional affiliation contributions

Fig. 3 shows the top 20 contributions per document publications of affiliated institutions in LULC studies. The University of Chinese Academy of Sciences was the most predominant institution, followed by the Institute of Geographic Sciences and Natural Resources Research and Beijing Normal University (n = 73 articles, 2.69 %), (n = 55, 2.03 %) and (n = 49, 1.81 %) published studies respectively. Chinese-affiliated institutions hold the top ranking in publications in scientific studies related to LULC, demonstrating their significant position in the field. Addis Ababa University in Ethiopia ranks fourth among the developing nation’s top 20 affiliations per document, with (n = 23, 1.55 %) publications. Unlike other developed nations, Ethiopia’s exceptional research output suggests that autonomous research funding strengthened its capacity (Fedderke & Goldschmidt, 2015). The study indicates that advancements in research on LULC may have increased the productivity of this field over the study period.

Close

Fig. 3

Top 20 institutional affiliations per document publications on LULC change studies.

Export to PPT

3.5 Land use and land cover change author's keywords

Table 3 shows the top-ranking author’s keywords (DE) and keywords plus (ID) in the LULC change research field, with 4499 and 8051 calculated based on occurrences and percentages. The list of the top 20 selected author keywords and keywords included approximately 12 common author keywords. “Remote sensing,” “GIS,” “Landsat,” “urbanisation,” “deforestation,” and “agriculture” are critical factors in understanding land use and land cover change, which influences “climate change.” It is worth noting that “remote sensing,” “climate change,” and “land use” were found among the top five authors’ keywords as well as in the keywords plus section. The results show that “remote sensing” (n = 235) ranked highest with most authors’ keywords (DE), while “land use” (n = 935) ranked first with most authors’ keywords plus (ID). Research on LULC change frequently recognises climate change challenges and emphasises the impact of these factors and LULC change responses on the process. The study highlights the impact of both natural and human-made factors on land cover, which can lead to irreversible loss and environmental changes. The interaction between land use patterns and climate change influences the structure of forests, grasslands, and ecosystems (Mndela et al., 2023). The keywords analysis revealed climate change as the third most frequently used term in authors’ keywords (2.9 %) and the fifth most frequently used term (5.8 %) in keywords plus. The study reveals that remote sensing is the primary term for analysing climate change and land use impacts on LULC change studies. Other factors of the author’s keywords (DE) include deforestation, urbanisation, land surface temperatures, and agriculture with low frequency, suggesting more research development in LULC change studies. The study revealed that the keywords plus (ID) were more frequently used than the authors’ (DE) in the survey.

3.6 Density visualization map of highest-ranking author’s keywords in LULC change dynamics

Fig. 4 shows the density visualisation map of the top 20 authors’ keywords frequently used in LULC change studies, highlighting their scientific contributions to research outputs. Zipf’s law was used to rank frequent authors’ keywords based on research subjects, directions, and niche trends, utilising the principle of the least effort to determine the highest occurrence of the lowest-frequency word (Madrid-Casaca et al., 2021). The predominant author's keywords displayed higher density (light yellow colour), while less frequently used authors' keywords showed lower density (dark purple colour). Moreover, the author’s keywords such as “remote sensing,” “land use and land cover change,” “climate change,” “urbanisation,” “Modis,” and “NDVI” significantly influence the field. Other significant variables like soil erosion, urban sprawl, population, evapotranspiration, cropland, forest transition, runoff, afforestation, and land management had a lower density of keywords. Satellite remote sensing applications are crucial in estimating and detecting land use/land cover changes using multitemporal Modis NDVI data in response to climate change and urban sprawl.

Close

Fig. 4

Density visualization map of top author’s keywords co-occurrence.

Export to PPT

3.7 Top global cited published documents on LULC change research development

Table 4 presents the most cited international studies on LULC changes, outlining data sources, techniques, and findings for each case study within the research development scope. Most top-cited studies utilise Landsat thematic mapper (TM), consisting of Landsat 4 and 5, with seven spectral bands ranging from 30 to 120 m resolution (Hong et al., 2011; Nurwanda et al., 2016). Studies indicate that maximum likelihood supervised classification outperforms unsupervised classification in accurately predicting LULC change (Kaur et al., 2013; Alloghani et al., 2020). The findings show that supervised classification techniques like maximum likelihood are commonly used to monitor and map land use and cover changes. Climate change, land surface temperature, urbanisation, human-induced degradation, and socioeconomic developments contribute to the long-term dynamics of land use and land cover (Salleh et al., 2013; Purwanto and Kurniawan, 2016). Intervention strategies, such as expanding urban green spaces and creating buffer zones around protected areas, aim to mitigate the severe impacts of land use and land cover changes (Buyadi et al., 2013). Policy decision-making and interventions supporting advanced remote sensing technologies are limited to LULC change research in the Global South (Qingqing et al., 2012; Dlamini et al., 2021). Recent environmental policy interventions are expected to advance, with cited articles in remotely sensed data in LULC changing research trends for future development and mitigation of land degradation (Table 4).

4.1 Limitation and future research prospects on land use and land cover changes

This study’s limitations include its focus on 1993–2022, which restricts its applicability to a broader context and other LULC change domains. The three databases used were chosen due to the early attention paid to LULC change studies in 1993, making it challenging to generate comprehensive datasets. The analysis was condensed to focus on significant research on LULC changes to accommodate the niche area. This suggests that further research using additional databases can help quantify the potential improvements and developments in understanding the relationship between external and internal factors influencing change dynamics. Further analysis is needed to understand the full implications of anthropogenic land use and carbon storage. This will aid in monitoring substantial investments in sustainable land management practices through national and global statistical infrastructure. The study suggests integrating remote sensing-based models like the future land use simulation (FLUS), the Land Change Modeler (LCM), the Cellular Automata-Markov Chain (CA-MC) model, and patch-generating land use simulation (PLUS) to predict future LULC changes. Modelling the effects of land use and land cover changes will provide a robust assessment of spatiotemporal characteristics, trends, drivers, and future predictions.